2024 – 2026 · Live · College, then rebuilt
Infrastructure Tracker — where every asset is, live
Where every asset is, who holds it, and what is about to fail. First built in college in 2024; rebuilt in 2026 as a self-hostable tracker for any organisation — campus, office, hospital, workshop — and deployed with a demo you can sign in to.
As posted
Detect · decide · actAn asset failing, in the wrong place, or about to fail — a warranty ending within thirty days, or two failures in ninety.
Who may act on it. Managers are scoped to the buildings they run; every server action checks the permission matrix before touching the database.
Opens a failure ticket, flags it in an audit, or reallocates it — and logs who did what, with the before and after.
The problem
Asset records lived in spreadsheets and in people's memory. Nobody could say where a given machine was, audits burned days of manual reconciliation, and failures were discovered by walking into the room.
The approach
One deployment is one organisation. Locations are building → floor → room, each with someone in charge; an asset carries tag, category, serial, status, custodian, cost, warranty and its full history. Stock is whatever is unallocated, by category, with low-stock thresholds. A failure is a ticket — open, acknowledged, in progress, resolved, closed — and the asset is under repair while one is open. An audit walks a room with a checklist and flags what is missing or misplaced. The dashboard's five numbers come from aggregation pipelines. Four roles — admin, manager, technician, viewer — and every mutation goes through a server action that validates, checks permission, writes, and appends to the activity log.